Current Position:  Home   >   Research   >   Invention Project   >   Content

National Natural Science Foundation project: The research of airport video emotional incident detection based on multi-level submerged condition

release time:2012-10-17

The research of airport video emotional incidentdetection based on multi-level submerged condition

Project number: 61272206

Main contents of the research:

Video emotional incident detectionis ahot issue which raised the attention from many scholars in the field of computer science. The reason is that it is a basic problem, relating many fields. What’s more, it is a kind of much-needed key technology. This project aims to create the applying conditions in random field model for video emotional incident detection and find the new method which have distinguish force characteristics . In particular, the research mainly intends to carry outthe following three aspects: first,overcoming the lack of monolayer condition random field which can't realize the deficiency of the multilayer reasoning, put forward multi-layer submerged random field with condition, and establish new methods for the detection of emotional events based on multi-level submerged condition of the random field. Second,facing the emotional incident detection, we will improve existing local binary pattern facial expressions recognition method based on local binarization mode.Among them, in order to overcome the weak point that the traditional facial expression recognition method cannot make use of the relationships between features, we will come up with facial expressions recognition methods based on the submerged condition with the random field; In order to overcome local binarization mode inadequate in describing facial movements,, the author puts forward alocal binarization mode with the facial features of the information together asa facial expression recognition characteristics. Third, we will study how to improve the edge histogram method, so that it can better detect facial features, as well as research to identify facial features state pedigree regression.

Address: Science Hall, Huazhong Normal University, 152 Luoyu Road, Wuhan, Hubei, P. R. China, postcode: 430079
(National Engineering Research Center for E-Learning)
Tel: 027-67867024 Fax: 027-67862995 Email:

visitor volume